Web Handling & Converting Blog

The Optimum Number of Defects

As strange as it may seem, a certain amount of (waste or delay) defects is desirable. Seldom is the optimum amount of any common pathology to be zero. To have a process or product so trouble free would indicate an economic overdesign of raw materials, processes, products or machinery. So much for zero defects or even Six Sigma, RIP.

Those interested in learning more about the how’s and why’s of the desirability of a certain level of defects can read further. The single most important and relevant article I’ve ever written, Optimization by Integrating Engineering and Business Models, has been given at two international conferences and can be downloaded from my website.

In Remission

Some of our web handling pathologies are almost completely curable. An example would be wrinkling due to roller misalignment. True, the medicine might be quite expensive; optical alignment of every single roller. The medicine might also be inconvenient; your mechanics may not be able to change rollers out without special care. Treatment will also take time; something like one hour of downtime for every single roller, yes idler rollers too. Finally, you may have to take the medicine for the rest of the life of the machine. Yet, unless you have unusual geometries or very stiff materials (such as metals), you can with commercial techniques get roller misalignment below the level necessary to cause misalignment wrinkling. This critical misalignment value can be calculated via the formulas in my Mechanics of Rollers book or much more conveniently with the TopWeb program.

Sorry to say however, is that most of our web handling pathologies are incurable. By that I mean that if you were ever were troubled by baggy webs, for example, you almost certainly will be forever. The same is true with misregistration, web breaks, most wound roll appearance issues (defects) and many, many more troubles. The way to think of most of our web pathologies is like herpes or cancer or alcoholism. The best you can expect is to be in remission.

A good colleague of mine recently said that he ‘fixed’ a longstanding problem in one plant that I had also worked on. However, in the very next sentence he admitted some new concerns (incidences) of the very same problem he claim to have ‘fixed.’ With all due credit, he almost certainly reduced the incidence significantly or perhaps even enormously, but that is not the same as eliminating it completely.

Winder Vibration

Winder vibration poses special diagnostic challenges way beyond any other web machine [1]. The reason has to do with the wound roll itself. The wound roll is a complicated element because it will have an imbalance and out-of-round much larger than any decent roller. Second, it changes rotational frequency during the course of the winding cycle due to increasing diameter. Thus the forcing function ‘sweeps’ out a range of frequencies rather than being single valued such as a roller. This means there is a much greater chance of exciting a resonance. Third, the wound roll can literally change shape during winding. Thus, for example, a roll can change from one sided (egg shaped) to two sided (oval) to three sided or more sided and/or back. This behavior again increases the chance of exciting resonances. Finally, the misshapen roll can excite vibration that pounds the misshapen roll further out of shape in a self-exciting fashion.

Another necessary concept is that of ‘bouncing’ grades [2]. There are certain grades that are notorious for generating wound roll vibration on almost any winder they are wound on and under almost any winding condition (though winding tighter is usually the best strategy). In paper, these grades include book, kraft, NCR, tissue and others. Sometimes a subtle change of chemistry, such as additives of a few pounds per ton, or a loading change on the paper machine’s calender can switch vibration on and off like a light switch. However other materials such as nonwovens and textiles also tend to vibrate if they are wound at higher speeds. What connects (most or) all of these materials together are two properties: high web-web COF and high ZD compressibility. To make a very long story short, these properties tend to make a wound roll get out of round due to self-excitation. A bump makes the roll bounce, and the bounce makes the bump worse etc.

The question will come up, “whose fault is it.” Despite the enormous complexities of the mechanics, this has a relatively simple test and answer [3]. Does the winder vibrate much when it is run at full speed, but without a wound roll? If so, the winder is clearly deficient. If not, the winder itself may be fine, especially if your product is known to be a bouncer, especially if other materials can be wound without such trouble. In that case, the wound roll is the exciting force, the wound roll is the primary mass in motion and the wound roll has the lowest spring rate of any member of the structure. All the winder is doing is going along for the ride, especially if it is a layon roller or rider roller moving in the direction that it must for roll build. This is one of the toughest problems in the web industry because it is so constrained. Almost the entire set of most important vibration factors belong to the roll product, which presumably can’t be changed, rather than the winding machine.

An analogy might be helpful. Consider a car and a road. Can the car run full speed? Yes, IF the road is straight and smooth. Here, the road is analogous to the wound roll and the car is winder. Those who wish to study further should consider the following classic articles.

1. Roisum, David R. Winder Vibration Can Reduce Operating Efficiency and Increase Maintenance. Tappi J., vol 71, no 1, pp 87-96, January 1988.

2. Olshansky, Alexis. Roll bouncing. TAPPI J, vol 80, no 2, pp 99-107, February 1997.

3. Roisum, David R. Tissue Winder Vibration. Perini J., vol 13, no 8, pp 33-36, August 1998.

Why Machines Vibrate

This is a simple question that has a simple answer; at least initially. Machines vibrate because they are forced to because of either roll(er) imbalance or nipped roll(er)s that are not round. The troubleshooter must identify the worst specific offending roller. This is easy with FFT (Fast Fourier Transform) vibration analyzers. The primary response of the structure is at a frequency equal to an integral number times the rotational frequency of the offending roller.

The response to these forcing functions is a bit more complicated. First, there is the concept of compliance or flexibility of a structure that includes the machine and its mounting (foundation). A machine structure would not vibrate much if it were quite stiff. Second, there is the possibility (or likelihood) of resonances. Here, a structure might vibrate 2-10X as much as it would otherwise when the offending roller’s frequency matches a particular system resonance, of which there are many. (Note: the idea critical speed of rollers has no meaning here or anywhere in the real world of web machinery. The idea is based on the erroneous assumption of infinitely stiff support.) However, all of these complications aside, a simple idea emerges for the troubleshooters. That is to identify the specific structural element that has the greatest relative displacement. This is often the member that is at an anti-node, but more accurately it is the member that has the greatest internal defection (and usually stress). This is the one element that might be eligible for stiffening on the next iteration of machine design.

How to Tape Rollers?

You can use any tape suitable for the application. Of course, simple masking tape is cheap and easy. However, masking tape is not very durable and is a $#&! to remove after it sits on a roller for more than a few hours. There are products that are more durable and easier to get off such as 3M 540 Traction Tape. Other tapes for other special situations would include cloth, duct, Teflon and so on. There is also a product called Printer’s Friend by Tesa that has a neat combination of properties that include air handling, release and traction. As you might expect, all of these products are far more expensive than the ubiquitous creped masking tape.

How you apply the tape is up to you but include annular, spiral with a pitch equal to the width of tape and spiral with gaps between the tape. One note on spirals: it does not spread. However, to just simplify the politics for people who believe without evidence, you can spiral outward from the center.

Why Tape Rollers?

Taping a roller is often used as a substitute for grooving or rubber roller covering. Tape can be a fast, inexpensive and flexible method to deal with certain challenges. It can also be a literal and figurative mess; causing waste, delay and confusion. The devil is, as they say, always in the details. There are four reasons you might want to tape rollers.

1. Changing speed ratio for geared systems. This application is quite rare and usually unintended, much like grinding a roller in a set and getting a speed mismatch.

2. To change the (web-to-roller coefficient of) friction for slippery products when slippage is otherwise occurring or to improve the release for tacky products that stick to rollers unacceptably or to increase roughness for the purposes of improving air handling or release.

3. Air handling can be improved greatly by applying a spiral of creped masking tape. This is useful when you have a smooth web running at hundreds of mpm that is slipping on the roller. Here it is not the spiral or the presence of a gap between the tape that does much of anything. The air handling primarily comes from the roughness of creped masking tape. A smooth rubber tape would not work as well even though the COF is higher. Spirals are entirely beside the point, it is just a convenient way to get the rough masking tape on. In fact, in modest cases you might regain traction by merely putting two annular rings on the ends of the roller.

4. You can get a modest concave spreader by putting a collar of tape at the edges of the web (only a couple of wraps please!). The advantage of using tape is that you can (in best practices) put it only when and where needed. While tape collars are not as neat as a concave cut into a roller, they are almost as effective and far more flexible. For example, you might put four layers when running a low modulus web such as vinyl, two layers when running a medium modulus such as PE, or one wrap when running paper and zero wraps when running foil. Note, for the 10,000th time; SPIRAL TAPING PROVIDES NO SPREADING.

There is one primary misuse of tape. Note for the 10,001 time; SPIRAL TAPING PROVIDES NO SPREADING.

Really Thick Webs

Are really thick webs subject to web handling laws? Of course. You can not break any laws of physics. However, the question is better posed as “are our current web-handling laws predictive for thick webs?” Here we are not so lucky as we were with strings. Thick webs are not necessarily a subset of thin webs. They are for certain problems. However, thick webs require immense traction to cause in-plane bending, such as with misaligned rollers, guides and spreaders. In most practical cases the traction is insufficient and thus the web handling theory must be able to handle the partial slip case (part of the web-roller width is in traction, but part is in sliding). Even more problematic is the bending stiffness about x and y axes. For example, bending stiffness in the y axis means that you have additional stresses due to the imposed curvature when the web goes around rollers and when wound around cores. This results in minimum roller and core diameters that can be calculated simply. However, other effects may not be so simple to calculate. Wrinkling theory is an area that is most upset by thick webs. However, you shouldn’t be wrinkling thick webs anyway. Shame on you. The resistance to wrinkling goes as the cube of thickness so there is no excuse.

String Theory

Is a string a web? In other words, does web-handling theory predict how a string will behave? The answer is yes in almost all cases. The reason is that a string is a subset of a web in the sense that the width or width to length span is effectively zero. Thus, any equation where zero (or at least very small) values for width or width ratio does not cause it to ‘blow up’ is predictive.

Because the effective width is zero, there is no resistance to in-plane bending. The web takes straight-line paths between (misaligned) rollers, guides and spreaders instead of the shape of beam bending, but this is already accommodated by web theory. What web handling theory that blows up is things like wrinkling and air entrainment because, quite practically, neither exists.

The Wide-Shallow or Burping Groove

The solution for the air bubble behind the winding nip described in the previous post was independently discovered in both the film and paper industries. A wide-shallow grooving had long been used in paper for reasons the designers could not articulate, but it prevented a problem we now know the mechanics of. In the film industry the groove may be called a ‘burping groove’ because it emulated one radical treatment: to momentarily lift the nip to let the bubble through. What happens with this groove is shown in figure above. The groove allows freedom for the wound roll layers to deform into the groove and thus form a tiny channel for the air to vent by the nip. The geometry of this groove is critical. Too much groove (width, pitch or depth) and you let too much air in. Too little groove and you have inadequate venting. The geometry that has been found to work is something like 1” wide and perhaps 0.010” deep by as wide of a pitch as practical. The geometry may be first simulated with tape before cutting the rollers or drum.

In the film industry, only one start and one revolution is required because every CD position in the winding roll is given a chance to vent on every revolution of the layon roller. Think of ‘rifling’ in a gun barrel and you get a rough idea of the geometry. However, in the paper industry a shallow angle such as this might cause tremendous noise and vibration at the much higher speeds they might run. Thus, in paper we use a more conventional pitch. The problem here, or in any case, is not the second or third nip (as in the case of a two drum winder). It is the inlet such as most layon rollers and the first drum of a two-drum winder. The more conventional pitch geometry in the paper industry lets in more air than one might like all the while it is venting it.

An (Air) Bubble in Front of a Winding Nip

Air bubbles can accumulate behind a winding nip as seen below. The mechanics of air entrainment modeling taught us that the air entrainment increases with: increasing speed, decreasing WIT (web tension and especially nip), increasing web smoothness and decreasing web caliper. What modeling did not do was to explain the potentially uneven distribution of air during winding nor describe what has become the de-facto treatment for the air bubble.

Why might air accumulate behind the nip? Simple theory says that the air should be uniformly distributed through the layers and across the width (with the exception of exhaust at the edges of the wound roll that takes some time). However, we know from ample experience that the air tends to accumulate in the low gage areas of the rolls, hence a CD position variation in air distribution. More importantly here, the nip can be an obstruction like a dam or squeegee and cause a rotational position variation in air distribution. Consider what would happen if only 99% of the air made it by/through the nip. That would leave 1% behind, being pushed ‘upstream’ under the first layer. On the next revolution another 1% might remain behind and so on in an accumulating bubble. Close observation will reveal that the air bubble may not be just under the outer layer but rather the outer several layers because this principle is not confined strictly to the outside wrap.

This process does not usually accumulate indefinitely. Because the air bubble upstream of the nip is at a higher pressure, it is even more motivated to get through the nip. Also, air escape on the edges may limit the response, especially for rolls that are narrower. In fact, there may be no accumulation at all. This response is extremely material dependent. The materials most prone to forming bubbles behind the winding nip are thin, have low modulus and have high tackiness. Many specialty films exhibit such a severe response that the operator might have to stop the winder, lance the bubble like a blister, flag the area as defective, and restart. However, at higher speeds, such as seen in the paper industry, we see this problem on two drum winders behind the drums and especially the rider roller even though paper possesses none of the risky web properties given above. Later the phenomenon was also noted and treated at the reel on the end of the paper machine.